postcards from mars
Downloading images from NASA's Mars Exploration Roversby Scott Tillitt
Stop what you're doing and look up, up to the heavens. Right now, perhaps, beaming through the ether on what must be the fastest wi-fi there is (the speed of light) are packets of binary code of what must also be the coolest vacation photos there are.
Since their airbag-cushioned landings in early and late January, NASA's two Mars rovers have been capturing panoramic images of the red planet's surface in their quest for clues of past life. Those images are sent the 120 million or so miles back to Earth three or four times a day to one of three radio telescopes in the US, Australia or Spain - with just a 10- to 11-minute download time - and then streamed to the Jet Propulsion Laboratory (JPL) in California.
The rovers' first images, the ones that show the landscape, each use a specially designed panoramic camera - Pancam - to explore the terrain before moving in for a closer look. The Pancam includes a pair of palm-sized, 9-ounce cameras (for stereo vision) with high-resolution color CCD (charge coupled device) detectors - similar to those found in most consumer-grade digital cameras, according to Dr. Justin Maki, a JPL imaging scientist on the Mars Exploration Rover (MER) project (NASA has an affinity for acronyms and abbreviations).
The Pancams each have 14 filters (plus two designed exclusively to observe the sun), including near-infrared, ultraviolet and the three primary colors: red, green and blue. The RGB filters are used to approximate "true" colors when processed back home.
The cameras sit on a bar on top of the rover's mast, allowing them to rotate a full 360 degrees for the panoramic view and swing up or down 180 degrees to capture elevation. The rover itself can take a 360-degree, stereo "Navcam" panorama of 20 images in about 20 minutes. But Pancam panoramas take several hours to a day or two, depending on the number of color filters requested. (Most of that time is spent storing each image in the camera's RAM and compressing the images to be sent back home.)
Raw images are released to the web in real time and have proven popular with the populace. But those images are then re-processed and stitched into mosaics as large as 4000 pixels high and 24,000 pixels around. The mosaics themselves can be generated in 20 seconds - larger ones can take tens of minutes - using a process called "ray tracing," which is similar to creating standard VR panoramas. The software, however, is not standard. It was written in-house at JPL by the Multimission Image Processing Laboratory (MIPL). No off-the-shelf stuff for NASA.
Since the eventual mosaics are created from many smaller images taken over time, each image may show changes in environmental conditions, dust levels for example, and lighting that have to be corrected. Even NASA scientists describe this as a "painstaking activity." This also involves approximating and compositing those RGB colors - which is nothing like color correcting, say, the shadow on a model's face or the tint of her lipstick. No one has been to Mars to know its true colors. Instead, the renaissance scientists use telescopic views, data and pictures from previous missions, and a touch of art.
The Pancams took four years to develop (the original concept dating back 10 or more) with a team that included a number of experienced imaging scientists and camera engineers -professional "space photographers," Dr. Maki calls them. Sounds like some expensive R&D, no? "I don't have any official numbers for the cost of the Pancams, but I can say that they were developed in parallel with the other MER cameras in order to minimize costs through economy of scale," he assures.
After four intense years, one can only imagine the excitement felt in the Jet Propulsion Lab when the first image arrived. It "was a very rewarding moment for everyone involved," Dr. Maki says in his matter-of-fact, science guy way. "That first image represented the work of not only the camera team, but also the 100-plus contributing engineers and scientists from all across the MER project."